Weighted mean of layers
weighted.mean.Rd
Compute the weighted mean for each cell of the layers of a SpatRaster. The weights can be spatially variable or not.
Usage
# S4 method for class 'SpatRaster,numeric'
weighted.mean(x, w, na.rm=FALSE, filename="", ...)
# S4 method for class 'SpatRaster,SpatRaster'
weighted.mean(x, w, na.rm=FALSE, filename="", ...)
Arguments
- x
SpatRaster
- w
A vector of weights (one number for each layer), or for spatially variable weights, a SpatRaster with weights (should have the same extent, resolution and number of layers as x)
- na.rm
Logical. Should missing values be removed?
- filename
character. Output filename
- ...
options for writing files as in
writeRaster
Examples
b <- rast(system.file("ex/logo.tif", package="terra"))
# give least weight to first layer, most to last layer
wm1 <- weighted.mean(b, w=1:3)
# spatially varying weights
# weigh by column number
w1 <- init(b, "col")
# weigh by row number
w2 <- init(b, "row")
w <- c(w1, w2, w2)
wm2 <- weighted.mean(b, w=w)